As Big Data takes hold, companies’ information systems are adapting to the ever-growing volume of digital data by setting up “data lakes”.
This very specific information storage solution is an opportunity for the finance department, provided its workings are properly understood.
Data Lake: a definition
As its name suggests, a Data Lake is a genuine lake of data.
A lake acts as a water collector, whether the water comes from rain, a river or melting snow. The water is kept without distinction or specific treatment, and remains permanently available for use.
A Data Lake does the same with information. It collects data from different sources, stores it regardless of its form (raw, semi-raw, refined) and keeps it available to the user.
Data Lake or Data Warehouse?
A common mistake is to believe that a Data Lake is simply an improved “Data Warehouse”.
While both systems do share the common goal of storing data, they are fundamentally different in their structure:
- A Data Warehouse is structured before storage, with each piece of data assigned to a well-defined part of the warehouse. In a Data Lake, data is selected after storage.
- Accessing the well-organised data in a Data Warehouse is relatively simple. By contrast, only an expert can extract data from a Data Lake.
- The structure of a data warehouse must be designed before it is set up, and can no longer be changed (or only with great difficulty). A data lake, on the other hand, is highly flexible and can adapt quickly to change.
These are therefore two completely different ways of storing information.
Different ways to set up a Data Lake
Setting up a data lake within a company is a major project that requires a number of strategic choices.
Data lake: on-premises or in the cloud?
A Data Lake can be implemented on the company’s premises or exist in the cloud. The choice is often based on the following criteria:
- Physical space: on-premises servers require room and dedicated infrastructure.
- Cost: cloud solutions require a much lower initial investment, but may prove more expensive in the long run
- IT skills: an on-premises installation requires the company to have experts on site, whereas the cloud solution makes it possible to delegate configuration, maintenance and upgrade tasks
- Data control: it is easier to control information kept on site than information stored in the cloud with a third party.
In general, micro, small and medium-sized businesses opt for a cloud solution, while large groups choose an on-premises installation.
The different architectures for a Data Lake
All data lakes share the same basic principle: storing information and processing it afterwards.
Today, 3 major architecture systems dominate the Data Lake market:
- Hadoop: a recognised and reliable “open source” solution whose main advantage is its price.
- Amazon AWS : a comprehensive range of products meeting a wide variety of needs.
- Microsoft Azure : a secure, high-performance solution that is easy to integrate.
Choosing the architecture best suited to the company’s needs requires an in-depth technical analysis and the involvement of external professionals.
The benefits of using a Data Lake
The structure of a Data Lake offers many benefits, which can be grouped into two broad categories.
Large data storage capacity
A data lake is the best solution for retaining a large quantity of data and thus ensuring that all the necessary information is available.
The absence of selection and processing prior to integration ensures that no data is discarded. It also allows faster, less resource-intensive storage, as it does not require significant computing power.
Ease of information processing
The raw information in a data lake can easily be used by analysts and by machine learning. Moreover, it remains permanently available in its original form, allowing simultaneous processing by several algorithms without the data being altered.
In practical terms, this means that a piece of revenue data can be used differently by the sales department, the finance department and the marketing department without any interference or loss of information.
The drawbacks of a Data Lake
As we saw above, a Data Lake is not an improved Data Warehouse but a different storage solution, and one that is not without its drawbacks.
Accessing the data you need
For some IT specialists, trying to access a specific piece of data in a data lake has replaced the expression “looking for a needle in a haystack”.
In itself, a Data Lake is useless without processing, analysis and selection tools that make it possible to obtain the desired information quickly and accurately.
Information confidentiality
The data held in a Data Lake comes from a variety of sources and from countries with different legislation. It is therefore difficult to ensure that a data lake complies with local laws on the confidentiality of information.
Likewise, “high-risk” information, such as financial data, may end up in a data lake without the necessary safeguards.
As an aside, the largest data leak to date was not caused by a group of hackers, but by a mistake in the use of a Data Lake.
Using a Data Lake properly
A data lake is an excellent storage solution, provided a few rules are followed. This prevents it from turning into a data swamp (Data Swamp)
- Acquiring the necessary skills. Setting up and maintaining a Data Lake cannot be improvised and requires a certain amount of experience.
- Involving the whole company, not just the IT department.
- Controlling data, to ensure its quality and reliability.
- Training in the IT tools used.
Would you like to find out more about the opportunities a data lake offers your business? You can contact Altermès’s specialists.
❓Do you have questions? Are you looking for more added value? More advice? Contact us so we can think through your situation together
👆 You can also browse our range of chartered accountancy services! We adapt to the size and complexity of your organisation.

